INDUSTRIAL AND ENGINEERING CHEMISTRY
May 1951
OF EXPERIMENTAL RESINWITH COMMERCIAL RESINS TABLEXIII. COMPARISON
Application formulation (shoe sole) Resin 21
*
e
Tensile 12'/315' F 24 Elon ation, % 127315 24 Shore A 12/315 24 Shore B 12/315 24 Crescent tear B 12 24 Ross flex, aged 24 hours, 212O F oven, 0.250inch sole 12/315 cycles Rating 24/315 cycles Rating Taber abrasion, H-22 wheels, 1000 grams, 500 rev., grams loss 12 min 24 min
Teat Formula D
A
B
C
1500 1675
1450 1525
1525 1400
1750 1600
290 305 90 88
300 275 83 83
390 395 84 84
70 67
63 63
147 156
173 133
250,046 5
287,587 1 5
E
F
G
1750 1475
1475 1425
1625 1700
1500 1375
370 340 85 86
320 360 87 85
250 370 87 86
430 405 72 73
305 315 84 84
64 64
65 66
67 65
67 66
52 53
64 64
137 143
162 141
175 180
156 172
125 94
169 178
65,000 201,643 250,046 10 10 2 160,000 287,587 287,587 10 5 2 5
0 411
0 368
0 275
0 432
0 400
0 321
90,061 250,046 250,046 250,046 10 3 6 4 9 287,587 42,210 196,560 166,000 5 5 10 10 10
0 405 0 399
0.302 0.336
0 386 0 373
0.252 0.303
0 321 0 361
FORMULATION GR-6 26 No. 1 smoked sheet Resin aa shown Ca1oi;rn silicate Zinc oxide Process oil
50.00 50.00 50 00 85.00 5.00 5.00
Cumar M H 21/n
Paraffin
Altax Methyl tuads Sulfur
Crescent tear A.S.T.M. D 62448, Die B Hardness, A.B.T.M. D 676421'. Olsen f l ~ w A.S.T.M. , D 569-48. Heabbftenmg Point* A 0*075 inch is placed over '/le-inch diameter mandrels, placed 2 inches on center and supports a 143-gram weight having a specified presser foot, In a circulation water bath, the temperature of which is raised a t the rate of 1 per minute. The temperatures are taken a t the firat deflection and when the sample is deflected 0.5 inch; the average of these temperatures is considered the heat-softening point.
8.
7.00 1.50 2.00 0.15 3.00 268.65
-
ACKNOWLEDGMENT
The authors wish to acknowledge the assistance of the Production Service Section of the Chemical Products Developnient Division, and the Chemical Engineering Division, for the preparation of the experimental polymers, and are grateful to the Goodyear Tire & Rubber Co. for permission t o publish this paper LITERATURE CITED
R E a s I V E D April
21, 1950.
Presented before the Division of Rubber Chemis-
try at the 117th hfeeting of the AVERICANCHEXICAL SOCIETY Detroit Mioh.
STUDIES OF CATALYST DISTRIBUTION
AND
Intrinsic viscosities, obtained by a conventional method, using 0.2 gram of resin in 100 ml. of benzene. Taber Abrasion. Using Taber machine, the sample was tested after 500 revolutions using an H-22 wheel and 1000-gram weight. Ross Flex. Ross flex, a standard of the shoe industry, was obtained on ,a 1 X 6 X 0.250 inch sample, cut with a l/lo-inch chisel awl perpendicular t o the length a t a distance of 2.5 inches from the attached end, aged for 24 hours in a 212" F. oven, and flexed on the Ross flex machine a t the rate of 100 flexes per minute through an angle of 90". The flex cut growth, measured in 0.1 inch, ia given as the rating after flexing.
(1) Border, Juve, and Hem, IND ENQ CHEM.,38, 955 (1946). ( 2 ) Cunningham, E. N., Rubber Age, 62, 187 (November 1947). (8) Holt, Susie, and Jones, India Rubber World, 121, 416 (January 1950). (4) Jones and Prqtt, Ibid., 117,609 (February 1948). (6) Sell and McCutcheon, Zbid., 119,66 (October 1948). (6) susie and Wald, Rubber 65, 637 (August 1949). (') Thies and Ibid.t 61f 51 1947)*
Coal Hydrogenation Catalysts
SOL WELLER'
1243
EngineeEing
I s:0Panc:r development -
M. G. PELIPETZ
U. S. B U R E A U OF MINES, BRUCETON, PA.
I
N T H E liquid phase hydrogenation of mal on a large scale, an
attempt is usually made to achieve good dietribution of the catalyst in the coal-oil paste. When the tin-ammonium chloride combination is employed, as in the Billingham plant of the Imperial Chemical Industries, Ltd., the tin is added in the form of tin oxalate, which is presumed to decompose under reaction conditions to give tin in a very finely divided state (3). Ferrous sulfate, when it was employed as a catalyst constituent in German plants at Politz, Blechhammer, and Gelsenberg, WN impregnated on the coal from aqueous solution (7). There have been almost no 1
Present address, Houdry Process Corp., Marcus Hook, Pa.
published reports, however, of the significance of catalyst dlstribution for coal hydrogenation, and apparently no systematic study has been made of this subject. Almost all batch autoclave investigations of solid coal hydrogenation catalysts have involved the addition of powered catalyst to the coal or coal-oil mixture (I,$, 443,IO). In the case of one bituminous coal, a comparisoa has been made of powered stannous chloride and stannous chloride impregnated on the coal (IO). Very little difference was observed between the modes of catalyst addition a t a concentration level of 1% tin; however, a t lower concentrations (0.5 and 0.1% tin) superior results were obtained with the impregnated catalyst. It
1244
INDUSTRIAL AND ENGINEERING CHEMISTRY
T h i s work was undertaken in order to learn more about the influence of the physical distribution of a catalyst on its effectiveness for promoting the hydrogenation of coal. The activity of a number of catalysts was found to be much greater when they are impregnated on the coal from aqueous solution than when they are added as powders to the powdered coal. The ball-milling of coal and catalyst together was of intermediate effectiveness. With a Wyoming coal, the activity of stannous chloride was equaled by impregnated niclcelous chloride and surpassed by impregnated ammonium molybdate plus sulfuric acid. If careful attention is paid to the method of catalyst distribution, the effectiveness of many coal hydrogenation catalysts can be significantly improved. This may permit the utilization of cheaper and more readily available catalysts.
has now been found that differences in catalyst activity caused by differences in distribution are t o be widely found in coal hydrogenation. The effect is occasionally very striking; in some cases, substances that are without catalytic effect when added as powders become equal or superior to the tin-halogen acid combination when impregnated on coal. APPARATUS AND PROCEDURE
h l l of the experiments were carried out in rotating I-liter autoclaves equipped with removable glass liners. The apparatus and procedures were identical with those previously described ( 9 , l O ) . All catalyst tests were made under the following “standard” conditions: 450” C. reaction temperature, 1-hour reaction time a t temperature, and 1000 pounds per square inch initial (cold) hydrogen pressure. Fifty grams of powdered coal were used in each experiment; in the ex eriments in which heavy-oil vehicle was added, 50 grams of veEicle Tvere used. T h e coal was impregnated with catalyst b immersing it in a n aqueous solution of the catalyst, removing txe bulk of the water by oven-drying a t 110’ C., and completing the drying in a n oven a t 50’ C. In those experiments in which the catal s t was ballmilled with the coal, the ball-milling was continuecffor 6 hours, I n the absence of either impregnation or ball-milling, the powdered catalyst was mixed with the coal only by tumbling in the rotary autoclave during the normal preheating period of 1.5 hours. MATERIALS
Rock Springs coal from the D. 0 . Clark mine No. 9, Superior, Wyo., and Bruceton coal from the Bureau of Mines experimental mine a t Bruceton, Pa., were studied. The heavy-oil vehicle was the product of a pilot plant run on the hydrogenation of Rock Springs coal; it contained 11.8% benzene-insoluble material, 39.3 yo asphalt (benzene-soluble, hexane-insoluble material), and 48.9% oil (hexane-soluble material). Ultimate analyses, on a moisture-free basis, of the two coals and the oil vehicle are shown in Table I.
TABLE I. ULTIMATE ANALYSES O F COALS AND Constituent
Rock Springs Coal, %
OIL VEIIICLE
Bruceton Coal,
Oil Vehicle,
%
%
CATALYST T E S T S
ROCKSPRIKGS COALTESTIN THE ABSENCE OF VEHICLE. Table I1 summarizes the results of catalyst tests with Rock Springs coal in the absence of vehicle. Under the conditions employed, when no catalyst is present Rock Springs coal shows the low liquefaction
Vol. 43, No. 5
of 33%. The term liquefaction is used here t o represent the conversion of moisture- and ash-free coal t o gaseous and liquid products, including asphalt. Addition of powdered ferrous sulfate (1% Fe) is almost without influence on the hydrogenation (run 1592). Impregnation of the coal with ferrous sulfate (1% Fe), however, causes a sharp increase in the liquefaction to 85%, with correspondingly high productions of asphalt and oil (run 1257). Part, though not all, of the potential catalytic effect of ferrous sulfate can be attained by ball-milling coal and catalyst (run 1357). Similar results are observed with nickelous chloride. Impregnated ferrous sulfate is as effective a t a concentration of 0.5% iron as at the higher concentration, but much poorer results are obtained with 0.1% iron. The catalysis is not shown by all ferrous salts; ferrous chloride and ferrous bromide are very poor catalysts even when impregnated. This may be due t o a poisoning effect of the halogen ion, as impregnation with ammonium chloride simultaneously with ferrous sulfate causes a great decrease in the catalytic activity of the ferrous sulfate (run 1568). The behavior of nickelous chloride is very similar to that of ferrous sulfate. Added as a powder, nickelous chloride (1%Ni) is a very ineffective catalyst. Ball-milling with the coal increases its activity, and impregnation results in an activity a t least as great as that of stannous chloride, previously considered t o be the coal hydrogenation catalyst par excellence. Decrease in the nickel concentration to 0.5% causes a perceptible drop in extent of hydrogenation, and a decrease t o 0.1% results in very poor catalysis. Impregnated nickelous sulfate is a moderately effective catalyst, but not as good as nickelous chloride. It is interesting to note that with iron, the sulfate has a higher activity than the chloride, whereas the reverse is true with nickel. Impregnated stannous chloride gives results similar t o those of impregnated nickelous chloride, the catalytic effect decreasing sharply as the concentration is decreased from 0.5 t o 0.1% tin. At a concentration level of lyo,however, tin is exceptional, and possibly unique, in being almost as effective when added as powdered stannous chloride as when impregnated. The same result was observed earlier with Bruceton coal (9). The relative effectiveness of the powdered catalyst is much less a t a concentration of 0.5% tin, whereas at 0.1% the catalytic effect is small for both powdered and impregnated catalyst. Ammonium molybdate added as a powder was reported to be noncatalytic for the hydrogenation of both Rock Springs and Bruceton coal (IO). Impregnation of the coal with ammonium molybdate, however, causes a striking increase in the extent of hydrogenation, liquefactions above 90% being observed (run 1436). German data (8) indicated that with low range (brown) coal, the catalytic effectiveness of ammonium molybdate is enhanced by neutralization of most of the alkaline ash of the coal with sulfuric acid. The coal in runs 1463 and 1523, Table 11,was treated in this way; enough sulfuric acid was added t o the aqueous solution of ammonium molybdate during impregnation to neutralize 90% of the “alkalinity” of the coal. The “alkalinity,,, defined as the weight of sulfuric acid neutralized by a given weight of coal, had been determined previously. Seutralization with sulfuric acid makes the effectiveness of impregnated ammonium molybdate even higher (run 1463). At a concentration of 1% molybdenum, a liquefaction of almost 95% is observed with Rock Springs coal under these conditions, along with an exceptionally low asphalt yield and an exceptionally high oil yield. The results are significantly superior to those obtained with stannous chloride a t the same concentration of metal; they are the best the authors have ever observed with Rock Springs coal. I n the absence of neutralization, the catalytic effect of impregnated ammonium molybdate drops sharply a t concentrations below O.5y0just as in the cases of ferrous sulfate, nickelous chloride, and stannous chloride. Neutralization (90%) of the coal with sulfuric acid, however, causes a large increase in the liquefaction produced by impregnated ammonium molybdate at a concentration of 0.1% molybdenum (run 1523), although the con-
INDUSTRIAL AND E N G INEERING CHEMISTRY
May 1951
version of asphalt to oil is not as well catalyzed. The combination of sulfuric acid and impregnated ammonium molybdate is, in fact, by far the best catalyst for Rock Springs coal that has been tested in this laboratory. Impregnation does not have universal effectiveness. As is to be expected, most materials that show no, or poor, catalytic activity when added as a powder are still poor catalysts when impregnated on coal. Among such materials are cobaltous chloride, zinc sulfate, and zinc bromide. I n the case of certain insoluble catalysts, intimate ball-milling of the catalyst with the coal does produce some increase in activity (compare the results with ferrous sulfate above). Lead chloride is still ineffective when ballmilled with the coal, but lead bromide, which is a fair catalyst even when added as a powder, produces a liquefaction of 80% on ball-milling. The data suggest that lead bromide would be a highly effective catalyst if it could be impregnated on coal. Ballmilling also produces a n increased activity in “copper chromite” (Adkins’ catalyst), but this material is still a poor coal hydrogenation catalyst even under these conditions. BRUCETON COAL TESTI N THE ABSENCEO F VEHICLE. The importance of catalyst distribution for catalytic activity has been verified in the cases of ferrous sulfate and nickelous chloride with Bruceton coal in the absence of vehicle. These experiments are summarized in Table 111. Ferrous sulfate added as a powder has no appreciable effect on the hydrogenation of Bruceton coal, and powdered nickelous chloride is even a negative catalyst. For both materials, impregnation causes a marked increase in catalytic activity, the liquefaction rising to over 80%. Impregnated
+
TABLE11.. EFFECTO F CATALYST DISTRIBUTION WITH ROCK SPRINGSCOAL (No vehicle)
Run NO. 1228
4
*
% of Moisture- and Ash-Free Coal Mode of Gaseous Distribu- Liquehydrotion.“ faction Asphalt Oil carbons 33.4 2.8 10.4 9.0
Catalyst
...
None
1592 1257 1357 1539 1535 1403 1402 1568
FeSO4 (1% Fe) FeSOr (1% Fe) FeS04 (1% Fe) FeSOa (0.5’7 Fe) FeSOa ( 0 . 1 % Fe) FeCh (1% Fe) FeBrz (1% Fe) FeSO4 1% Fe) 0.5% NH&l
P I BM I
1552 1336 1442 1563 1494 1321
I I
38.9 84.9 66.2 85.0 53.0 44.8 50.4 52.1
6.9 38.9 24.1 42.8 17.7 4.8 7.7 14.8
8.1 21.7 21.2 19.8 10.4 12.2 17.9 13.8
13.1 15.0 11.4 13.0 13.7 15.2 12.4 13.7
NiClz (1% Ni) NiClz 1 Ni) NiClz NiClz 0 . 5 Ni) NiCh ( 0 . 1 Ni) NiSO4 (1% Ni)
P I BM I I I
44.2 88.3 56.9 81.7 47.5 78.9
6.8 15.5 17.2 36.8 8.3 33.0
13.2 45.3 19.3 21.7 15.9 14.8
13.4 18.0 10.4 14.0 12.0 10.0
1598 1581 1393 1441 1391 1509
SnCh (1% Sn) SnClz (1% Sn) SnClz ( 0 . 5 % Sn) SnClz ( 0 . 5 % Sn) SnCln 0 1 % Sn) SnCln t o : 1% Sn)
P
P I
I P I
82.3 88.3 67.9 86.3 52.3 54.6
26.5 19.9 23.4 37.3 17.1 13.8
29.2 41.4 20.0 28.0 11.4 14.0
14.5 15.5 13.4 11.2 12.5 15.7
1121 1436 1463
A M b (1% Ma) AM 1 7 Mo) AM Ma), neutralize% AM ( 0 . 5 7 Mo) AM ( 0 . 2 q Ma) AM ( 0 . 1 2 Mo) AM ( 0 . 1 % Mo), neutralized
P I I
33.7 92.7 94.7
1.0 27.2 10.0
13.8 41.1 62.6
8.1 13.6 13.9
I I
89.4 61.3 53.4 85.1
28.4 25.9 14.9 42.0
37.7 11.1 14.9 18.8
13.8 14.7 12.7 14.4
1310
CoClz (1% Co)
I
58.0
11.2
16.3
8.3
1314 1389
ZnSOa (1% Zn) ZnBrz (1% Zn)
I I
49.5 54.0
12.5 10.6
14.5 21.9
11.8 14.6
1292 1361 1594
PbClz (1% Pb) PbBrz (1% Pb) PbBrz (1% Pb)
BM BM P
45.3 80.7 71.1
7.0 27.3 28.6
15.5 28.5 20.7
12.3 15.5 12.6
1446 1518 1493 1523
+
Ilgr
[lq
1593 Copper chromite (2%) 1394 Copper chromite (2%) P , powdered catalyst added to charge; I, catalyst impregnated on b AM, ammonium molybdate.
I I
I I
P 42.2 7.4 9.1 13.6 BM 65.1 24.2 17.3 12.5 charge; BM, catalyst ball-milled with charge.
1245
OF CATALYST DISTRIBUTION WITH BRUCETON TABLE 111. EFFECT COAL (No vehicle)
% of Moisture- and Ash-Free Coal Run No. 1623
Catalyst FeSOr (1% Fe)
1245 1236 1259 1453
FeSOr FeSOa FeSO4 FeSO4
1% Fe) { l %Fe) ( 1 7 Fe) (1% Fe)
Mode of Distributiona
. ..
P I BM Aq
Liquefaction Asphalt 46.9 15.1 55.6 81.6 78.4 68.4
27.1 48.7 43.7 36.1
Oil 10.4
Gaseous hydrocarbons 12.0
14.2 16.7 18.1 13.5
10.4 12.1 12.2 15.2
NiCln (1% Ni) 12.0 11.4 P 40.0 11.1 NiClz (1% Ni) I 82.6 36.7 15.0 28.3 ’P, powdered catalyst added t o charge; BM, catalyst ball-milled with charge; I, catalyst impregnated on charge; As, aqueous solution of catalyst placed in autoclave with charge. 1158 1362
TABLEIV. EFFECTOF CATALYSTDISTRIBUTION WITH ROCK SPRINGSCOAL
Run 1506
Catalyst None
1469 1480
AMC (1 0 Ma) AM ( 1 z M o )
NO.
(Heavy oil vehicle) Yo of Moisture- and Ash-Free Coal Mode of Gaseou s Distri- LiquehydroOilb carbonsb butiona factionb Asphaltb , . 64.9 47.0 -4.6 10.4
.
P I
82.2 93.0
76.8 40.4
-17.7 31.8
12.7 9.4
NiClz 1% Ni) P 39.4 13.5 4.0 10.2 NiClz { l % Ni) I 82.6 19.5 44.4 9.2 a P, powdered catalyst added t o total charge: I, catalyst impregnated on coal. 6 Corrected for contribution of vehicle. C AM, ammonium molybdate. 1486 1526
nickelous chloride is superior to ferrous sulfate for asphalt reduction, just as in the case of Rock Springs coal. Ball-milling of ferrous sulfate with the coal again results in a significant improvement, compared with the powdered catalyst, but not as much as is obtained with impregnation. An attempt t o impregnate the coal in situ during hydrogenation gave rather poor results (run 1453). ROCKSPRINGSCOALTESTIN PRESENCE OF VEHICLE. The experiments reported in Tables I1 and I11 demonstrate the importance of catalyst distribution under conditions when no oil vehicle is used. It was conceivable that a vehicle, if present, would distribute even powdered catalysts so well that the effects observed in the absence of vehicle would disappear. T o investigate this point, several runs were made on the hydrogenation of Rock Springs coal in the presence of a coal hydrogenation heavy oil; the results are summarized in Table IV. The data given for product distributions have been corrected for the amounts of organic insolubles, asphalt, oil, and gas obtained when the vehicle alone is hydrogenated under the same conditions. As this procedure assumes that the hydrogenation of vehicle proceeds identically, whether or not coal is present, the corrections used can only be considered as approximate. Comparison of run 1506 with run 1228, Table 11,shows that in the absence of catalyst, a much higher liquefaction is obtained with Rock Springs coal in the presence of a vehicle than in its absence. This may be a real effect of vehicle or it may be due to the presence of traces of tin catalyst in the vehicle arising from the manner in which the pilot plant operation was carried out. The negative values for oil in runs 1506 and 1469 originate from the correction procedure discussed above. I n addition, because of the empirical nature of the asphalt and oil analyses, the values for oil tend to be low whenever the quantity of asphalt is exceptionally large. This effect is especially great for run 1469. I n the presence of vehicle, the effect of catalyst distribution is qualitatively similar to that in the absence of vehicle, but there are quantitative differences. Powdered ammonium molybdate is appreciably catalytic (run 1469); even in this case, however, impregnation on coal markedly increases the liquefaction and in-
I N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
1246
creases the conversion of asphalt t o oil. Nickelous chloride gives striking results: Added as a powder, it behaves as a negative catalyst; impregnated on coal, i t increases liquefaction and especially the asphalt-oil conversion. It is clear that although the presence of vehicle changes the quantitative results obtained, catalyst distribution is still of major importance. DISCUSSION
These experiments emphasize the role of catalyst distribution in determining activity. The more intimate is the contact of the catalyst with coal, the greater is the catalytic effectiveness; in fact, changing the mode of distribution in some cases changes the level of catalyst activity from a negligibly low to an extremely high value. I t is, of course, not surprising that the intimacy of contact of catalyst with substrate should be of great importance, but it is a factor that has been a t least tacitly overlooked in many atudies of coal hydrogenation catalysts. This work has several important implications. I n the first place, for highest catalytic activity the best possible catalyst distribution should be sought. Unless experiments on a given coalcatalyst combination prove contrary, i t would appear preferable either to impregnate the coal with catalyst from solution or, possibly, t o spray the powdered coal with catalyst solution. The latter alternative would be more feasible in a large plant. This can be done, of course, only for catalysts that are soluble; in other cases extended ball-milling of coal and catalyst presents another, but not m effective, possibility. Another result is that, for some coals, a t least, the unique position of tin compounds as hydrogenation catalysts no longer exists. With Rock Springs coal, for example, impregnated nickelous chloride is as good as impregnated stannous chloride, and impregnated ammonium molybdate plus sulfuric acid is
EngFnyring
Vol. 43, No, 5
definitely superior. Tin compounds are still peculiar in that, at high concentrations, they are almost as effectjive in powdered form as impregnated on the coal. However, this may be takcri t o mean that although a special mechanism, as yet not understood, exists that permits good distribution of tin compounds no mattw how they are added, no special or unique character need he attributed to tin as a catalyst for the coal hydrogenation reaction proper. The mystery surrounding tin is thus pushed back OIW step, and it becomes less important for understanding the chemistry of coal hydrogenation. ACKNOWLEDGMENT
The authors acknowledge the valuable assistance of Sam Friedman, James Bayer, Louis Frank, and Robert Salmon in obtaining the data reported in this work. LITERATURE CITED
(1) Abe, R., Huzikawa, S., Kakutani, T., and Okamura, T., J . SOC. Chem. I d . J a p a n , 41, supplementary binding, 417 (1938). ( 2 ) D e p l . Sci. Ind. Research (Brit,), Fuel Research, “Rept. for Period Ended March 31, 1932,” p. 54; “Kept. for Year Ended March 31, 1933,” p. 100. (3) Gordon, K., J . I m t . Fuel, 20, 42 (1946). (4) Hlavica, B., Brennstoff-Chem., 9, 229 (1928). (6) Kurokawa. M.. J . Soc. Chem. I n d . Janan. 45. 1033 (19421 (6) Kurokawa, M., Hirota, TV., Fujiwara, K., and Asaoka, h , J . Fuel Soc. J a p a n , 18, 31 (1939). (7) Sherwood, I?. W., Dept of Commerce, OTS R e p l . , FIAT 952
(1947).
(8) Technical Oil Mission Reel, 75, frames 368-9. (9) Weller, S., Pelipetz, M. G . , and Friedman, S., IND. ENG.CHEM., in Drew. (10) Well&, S., Pelipetz, M . G., Friedman, S., and Storch, H. H., Ibid., 42, 330 (1950). RECXIVED September 28, 1950.
rization
pOCeSS
development
I V. C. NEKLUTIN, C. B. WESTERHOFF,
AND
L. H. H O W L A N D
U N I T E D STATES RUBBER CO., NAUGATUCK, CONN.
T h e development of sugar-free redox recipes for the polymerization of GR-S at 41“ F. was undertaken in order to obtain simpler and more economical systems. T h e activatorsstudied in these recipes were: ferrous SUIfate-potassium pyrophosphate complex which, however, still required a complicated make-up procedure; ferrous sulfide colloidal dispersion, in which the s a d e serves, not only to form a slightly soluble iron compound, but also to regenerate the ferrous iron as it is oxidized; ferrous soap dispersion, which is prepared in situ and works well only when the emulsifier is a fatty acid soap; ferrous silicate colloidal dispersion which works best with rosin-type emulsifiersand results in the simplest and most economical iron recipe hitherto developed; and polyethylene polyamines which have an advantage in ease of operation. In the manufacture of synthetic rubber, these recipes afford simpler operation, better uniformity, economy, and freedom from pH drop on latex storage. Advantages to the
consumer are better uniformity of product at equal quality as compared with polymers from sugar-type recipes, particularly in carbon black master batches prepared from the latices.
A
S EARLY as 1941, workers in the field of synthetic rubber recognized improvements in the physical properties of copolymers of the GR-S type prepared by polymerization of butadiene-styrene a t low temperatures (1). However, at that tirncb, the existing recipes were too slow for economical use. Conscquently, since 1945,much work has been carried out by the vanous companies and universities in cooperation with the Office oi Rubber Reserve to develop rapid polymerization recipes for use at 41” F. or lower. The first fast polymerization redox recipes were patterned after German recipes using benzoyl peroxide as the catalyst. However, this material was expensive and resulted in a polymerization with a rapid initial reaction surge which mad(& temperature control rather difficult. A t this time, the Hercules
